HUMAN INPUT: A Sound/Visual Installation about how AI learns
About
Does A.I. Music have a future without us?Step into a physical simulation of how AI models learn to create music. In the space, a camera counts your presence as a unit of training data for a machine learning model. An empty frame produces only static—the random noise of an untrained system. As more people step into view, the simulated model scales up in complexity, producing music that sounds more and more human-like.
At a time where machine output is quickly becoming the norm, how much of our own creativity are we willing to hand over to algorithms?
This installation is both an educational insight into machine learning and a reminder that AI systems don't create anything on their own: without human input, there is no output.
We recommend allowing 15–20 minutes to experience the installation.
----------------
Created by:
Lyndon Riggs (Producer & Creative Direction) – BMus Student UniMelb, BSc (Physics)
Jonathan Latti (Technical Lead & Creative Direction) – UniMelb Alumni, Master of IT (Artificial Intelligence)
Supported by:
Melbourne Conservatorium of Music ‘IgniteLab’ Student Grant.
Date
Thursday 3 September 2026 6:30 PM - 8:30 PM (UTC+10)Location
Activity Room G01, Building 189
Bldg 189/761 Swanston St, Parkville VIC, Parkville Victoria 3010